Method and Apparatus for Operating an Automation System and Accounting for Concept Drift

1. A method for operating an automation system comprising: a) providing a learning-based prediction model for the automation system trained by process data comprising context of an automation process, b) receiving information about current context of the automation process, c) verifying context change by comparing the current context to the context of said process data, d) in the case of any context change verifying a concept drift by comparing pre-drift process data and post-drift process data, e) in the case of any concept drift re-training said learning-based prediction model with post-drift process data, f) in the case of no context change testing for random concept drift not detected by verifying context change, g) in the case of any random concept drift extend the current context by using data comprising previous context changes, wherein the current context is automatically extended on the basis of the context's influence on estimates of processing times, the influence being due to at least one dependency in the automation process, h) in the case of no concept drift and no random concept drift, using the provided learning-based prediction model.

2. The method according to claim 1 , wherein if such a random concept drift is detected without context change, the current context is extended with at least one adjacent triple stored in a context knowledge base, wherein the at least one adjacent triple is a triple <s; p; o> where s is a subject of the triple, p is a property of the triple, and o is the object of the triple, and wherein s or o is tracked in the current context.

3. The method according to claim 2 , wherein the context knowledge base comprises a context knowledge history.

4. The method according to claim 1 , wherein context is represented by features of a piece of product and/or automation process steps and/or in the automation process used apparatus.

5. The method according to claim 1 , wherein the model keeps track of a specified subset of a set of context.

6. The method according to the preceding claim 5 , wherein the initially tracked, specified subset only contains context instances that are directly associated with process data used for its training.

7. The method according to claim 1 , wherein the prediction model represents a fault prediction model.

8. An apparatus for operating an automation system comprising: a) means for obtaining a learning-based prediction model for the automation system trained by process data comprising context of an automation process, b) means for receiving information about current context of the automation process, c) means for verifying context change by comparing the current context to the context of said process data, d) means for verifying a concept drift by comparing pre-drift process data and post-drift process data in the case of any context change, e) means for providing post-drift process data for re-training said model in the case of any concept drift, f) means for testing for random concept drift not detected by verifying context change in the case of no context change, g) means for extending the current context by using data comprising previous context changes in the case of any random concept drift, wherein the current context is automatically extended on the basis of the context's influence on estimates of processing times, the influence being due to at least one dependency in the automation process, h) wherein, in the case of no concept drift and no random concept drift, the provided learning-based prediction model is used.

9. The apparatus according to claim 8 , wherein the apparatus comprises a context knowledge base.

10. The apparatus according to claim 9 , wherein said context knowledge base comprises a context knowledge history.

11. The apparatus according to claim 8 , further comprising: a) means for providing a learning-based prediction model for the automation system trained by process data comprising context of an automation process, b) means for receiving post-drift process data for re-training said model in the case of any concept drift.

12. The apparatus according to claim 8 , where in the prediction model represents a fault prediction model.

13. A computer program product, comprising a computer readable hardware storage device having computer readable program code stored therein, said program code executable by a processor of a computer system to implement a method directly loadable into the internal memory of a computer, comprising software code portions for performing the steps of claim 1 when said computer program product is running on a computer.